Effects of failure criteria on the constant humidity test results

Accelerated life-testing is much used in the reliability study of electronics. Failure criteria need to be assessed due to the application. For some applications even small changes in a variable used as failure criterion turns out to cause catastrophic failures while in another application even great fluctuation does not disturb the normal operation of the device. When the behaviour of the variable used as failure criterion is monitored prior to failure, regularities can be seen and thereby the starting failure may be predicted. In this work the effects of failure criteria was studied in humidity tests with three different test lots. Flip chip components on FR-4 substrates were used, and two of the test lots were coated with either epoxy or parylene C. The results show differences in the behaviour of daisy chain resistance between different test lots having different coatings causing differences in the reliability results with the three selected failure criteria.